Loading Academy Hub...
Scale Power BI to hundreds of millions of rows: Incremental Refresh configuration, `RangeStart` and `RangeEnd` parameters, historical partition freezing, and real-time Hybrid Tables.
[!NOTE] Refreshing a 500-million row model every morning via full data reload consumes hours of network bandwidth and hits gateway timeouts. Incremental Refresh partitions the table into immutable historical segments and refreshes only recently changed partitions.
RangeStart and RangeEnd.[TransactionDate] >= RangeStart and [TransactionDate] < RangeEndLastModifiedDate column.RangeStart and RangeEnd (strict DateTime type) on a single filtered date column (>= RangeStart and < RangeEnd — watch out for Query Folding!). Once published to Power BI Service, Incremental Refresh splits your 5-year table into static historical partitions (5 years untouched) and refreshes solely the rolling 7-day hot slice (10x faster refreshes!).Hot DirectQuery + Warm Import): On Premium / Fabric capacities, toggle "Get the latest data in real time with DirectQuery" inside Incremental Refresh so historical years stay blazing fast in Import VertiPaq while today’s live partition queries source tables in real time!Dev -> Test -> Prod) + XMLA Endpoints: Promote .pbip / Fabric Git-tracked semantic models safely via Deployment Rules switching parameter server/DB strings without manual Power BI Desktop re-uploads, and use Tabular Editor / XMLA to refresh single corrupted partitions selectively.Review the production code, company interview Q&As, incident traces, and comprehension quiz inside this lesson.
let
Source = Sql.Database("sql-lakehouse-prod.database.windows.net", "EnterpriseDW"),
dbo_fact_sales = Source{[Schema="dbo",Item="fact_sales"]}[Data],
// Mandatory Incremental Refresh Filter (Must fold into SQL WHERE clause)
FilteredRows = Table.SelectRows(dbo_fact_sales, each
[TransactionDate] >= RangeStart and [TransactionDate] < RangeEnd
)
in
FilteredRowsMicrosoft.PowerBI.DataMovement.Pipeline.Diagnostics.GatewayPipelineException: Operation timed out after 02:00:00.